Skip to content

About

A collection of AI agent skills to pull climate, elevation, and soil data at farm and trial sites. Part of manvia.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

33 Commits

Folders and files

Repository files navigation

ProjStatus Lifecycle License

Part of maniva

Environment data for on-farm trials and breeding

Brayden Youngberg, Kauê de Sousa

Daily weather, elevation and soil properties for points such as farms, trial sites and survey households, as agent skills and the code behind them. District lookups and climate hazards are planned. Built for breeders and the Climate Action team who need to know what a site or season is like, with tricot trials or on their own.

This is the environment layer of maniva, the AgrDataSci open commons of evidence, protocols, data and analysis for agricultural research. It is built from template-repo-data-analysis and works on its own.


Sources

Source What Resolution Access
CHIRPS v3 Daily rainfall, 1981 onward; latest date depends on store, 60°S to 60°N 0.05° public, no key
CHIRTS-ERA5 Daily Tmax and Tmin, 1981 onward; latest date depends on store, 60°S to 70°N 0.05° public, no key
AgERA5 Daily temperature, rainfall, radiation, humidity, wind, reference ET, 1979 onward; latest date depends on store 0.1° free Copernicus CDS key
Copernicus DEM Elevation 90 m public, no key
SoilGrids 2.0 Soil texture, carbon, nitrogen, pH, CEC, bulk density at six depths 250 m, or 1 km mean public, no key

Each source keeps its own licence and terms of use.


Skills

One folder per skill under skills/. Each has a SKILL.md an agent reads and a scripts/fetch.py it runs. Every script takes points as a CSV with id, lat, lon and writes parquet (or CSV if the output name ends in .csv), one row per point and day for weather and one row per point for elevation and soil, so weather outputs join on id and date, and elevation/soil outputs join on id. An agent builds that CSV from whatever form the user's points come in. Each script's fetch() function also works inside a Python pipeline, and each SKILL.md documents the source well enough to write your own reader, e.g. in R.

Skill What it does Status
chirps-rainfall Daily rainfall for points from CHIRPS v3 ready
chirts-temperature Daily Tmax and Tmin for points from CHIRTS-ERA5 ready
agera5-climate Daily weather for points from AgERA5 ready
elevation Elevation for points from the Copernicus 90 m DEM ready
soilgrids-soil Soil properties at standard depths for points from SoilGrids 2.0 ready
admin-boundaries District and province for points, or points in a named area, from GAUL 2025 planned
climate-hazards Dry spells, hot days, heavy rain and other hazards from any daily series, per growth stage or season planned
future-climate How those hazards shift under CMIP6 scenarios, from CCKP planned

Install

Scripts need only uv, which fetches Python and the packages on first run.

Claude Code. Symlink or copy the skill folders into ~/.claude/skills/ for every project, or into a project's .claude/skills/ for that project alone.

git clone https://github.com/AgrDataSci/environment-data.git
mkdir -p ~/.claude/skills
ln -s "$PWD"/environment-data/skills/* ~/.claude/skills/

Codex and Gemini CLI. Same, into ~/.agents/skills/ or ~/.gemini/skills/.

Without an agent. Run any script directly, e.g. uv run skills/chirps-rainfall/scripts/fetch.py points.csv rain.parquet --start 2020-01-01. When each point has its own date, such as a planting date, fetch a window per point: --date-col planting_date --days 100.

R users. Run the scripts as above, then read the parquet outputs in R with nanoparquet::read_parquet(), a small package with no dependencies, or arrow::read_parquet() if you already use arrow.


Network

The scripts read remote stores and uv installs their packages on first run. Sandboxed agents, such as a cloud chat session, and some institutional networks block those hosts. python3 scripts/check.py lists what is blocked, using the standard library only. The hosts are in SKILL.md, with what an agent does when they are out of reach.

Tests

uv run tests/test_fetch.py   # offline, synthetic stores
python3 tests/smoke.py       # two points from every live store

The offline tests run on every pull request. The live test runs on pull requests that change skills/ or tests/, and every Monday to catch a store that moved.


Cite

Cite with CITATION.cff.

License

Code is released under the MIT License, with rights held by the International Plant Genetic Resources Institute (Bioversity International). Documentation and protocols are CC BY 4.0. Third-party data keep their original licences.

About

A collection of AI agent skills to pull climate, elevation, and soil data at farm and trial sites. Part of manvia.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages